Biogéographie deMadagascar, 1996 : 73-83
MALAGASYLENDO-AUSTRALO-MALESIANPHYTOGEOGRAPH" CONNECTIONS
George E. SCHATZ
Missouri Botanical Garden, PO. Box 299, St. Louis, Missouri, 63166-0299, U.S.A.
USTRACT.- Despite the continuous close proximityof Madagascar to Africa sinceits separation from the continent ca. 165 MXA, the Malagasy flora exhibits a remarkably high affinity with the Indo- autralo-malesian florasfar to the east. Such phytogeographic connectionsare especially prevalent among easternhumid forest taxa, and representboth ancient vicariance that hasresulted in relictual (Cretaceous) Gondwanan disjunctions, as well as continuous dispersal events across the Indian Ocean. Threemajor patterns of dispersaVvicariance modality can be identified: 1) Cretaceous dispersal to Madagascar with ensuing distributions from India (andor South Africa) across Antarctica to South America and Australo-E. Malesia during the timeof the initial radiation of the angiosperms; 2) Eocene- Oligocene (and continuing to the present) dispersal to Madagascar (and Africa) from Laurasia and W. Malesia viaIndia @re- and post-collision with Asia) alongt( Lemurian Stepping-stones)) in the western IndianOcean; and 3) continuous(and recent) long distance dispersal (LDD) to Madagascar as a function of the prevailing easterlywinds and Indian Ocean currents.
KEY-W0RDS.- Madagascar, Indo-Australo-Malesian, Phytogeography, Vicariance, Dispersal
RESUME.- Malgré la proximité de Madagascar et de l'Afrique depuis leur séparation,il y a quelques 165 millions d'années, la flore malgache montre unefinité remarquable avec celles dela région Indo- australo-malésienne très éloignéevers l'est. Ces relationsphytogéographiques sont particulièrement fréquentes dans les taxons des forêts humides de l'est de Madagascar; elles représentent d'unepart une vicariance ancienne qui a entraîné des disjonctions reliques gondwaniennes (Crétacé) et d'autre part une série continue de dispersionà travers l'Océan Indien. Trois modèles principaux de dispersionlvicariance à Madagascarpeuvent être distingués : 1) Au Crétacéune dispersion qui a entraîné des types de répartition qui s'étendent de l'Inde (et/ou l'Afrique du Sud) jusqu'à l'Amérique du Sud et le région australo-est-malesienne àtravers l'Antarctique,pendant la périodeinitiale de diversification des angiospermes ; 2) Pendant 1'Eocène-Oligocène une dispersion encore actuelle et également africaine, de taxons Laurasiens et ouest Malésiens via l'Inde avant et après son entré en collision avec 1'Asia par l'intermédiaire d'îles (t< Lemurian Stepping-Stones B) situées dans l'ouest de l'Océan Indien ; 3) Une dispersion à longuedistance continue et plus récente due aux vents et auxcourrants océaniques dominants de l'est.
MOTS-CLES.- Madagascar,Région Indo-australo-malésienne, Phytogéographie,Vicariance, Dispersion
INTRODUCTION
As a former piece of the African continent that has remained in close proximity throughout its evolutionary history, Madagascarwould reasonably be expected to
In: W.R. LOURENçO (éd.) Editions de I'ORSTOM, Paris 74 G.E. SCHATZ possessbiotaa whose closestaflinities are African.Recognition therefore of phytogeographic connections with floras far to the east has attracted the interest of botanists since the classic work of PERRTER DE LA BATHIE (1936). Subsequent workers have adopted and refinedthe characterization of floristic af€inities by percentage generic overlapwithin arbitrarily delimited geographic units (tt elements B), amongwhich CRETACEOUS GONDWANAN RELICTS Paleomagnetic data from the westernIndian Ocean has firmly established the paleogeographic history of Madagascar (RABINowITZ et al., 1983). Figure 1 depicts a schematic representation of Gondwanan land masses fiom the time Madagascar arrived at its current position 121 MYA until the separation of Greater India fiom Madagascar 88 MYA(STOREY et al., 1995), bywhich time the break-up of virtuallyall of the individualGondwanan land masses had begun. Throughout the periodrifting was occurring both between AfricdSouth America and Greater Indidhtarctica, although eachpair remained within close proximitythrough the end of the Cretaceous/early Tertiasy. By 80 MYA, direct land routes between New Zealand and New Caledonia and AntarcticdAustralia had been severed (MILDENHALL, 1980); very slow rifiing between AustraliaNew Guinea and Antarctica had also begun, but New Guinea did not emerge above sea level until the late Eocene at which time Gondwanan elements first entered (AXELROD& RASEN, 1982). A dispersal pathway through Antarctica to South America remained possible untilthe close of the Eocene (RAS?EN, 1979). Thetime period spanned in Figure 1 isalso contemporaneous with the initial radiation of the angiosperms. Insofar as Madagascarbegan to separate from fica during the late Jurassic (165 MYA), it is probable that angiosperms were not present on the island during mostof its slow raRing to its current position. Paleobotanicaldata from Madagascar at this critical time period are scant. Palynomorphs fiom mid-Cretaceous (middle Albian to late Cenomanian) samples near Ahtsiranana belong to the Gondwana microfloral Province, dominated by bisaccate gymnosperms and ferns, but are lacking angiosperms (HERNGREEN et al., 1982).Another Cenomanian assemblage fiom Ankanotra is dominated by the conifer Classopollis; however, six different angiosperm pollen types are present. Despite the lack of late Cretaceous data, it is reasonable to assume that by the close of the Mesozoic, numerous angiosperm lines had established themselves in Madagascar via either direct land routes through India from the south, or by short distance dispersalacross water gaps. PHYTOGEOGWHIC CONNECTIONS 75 Whatneobotanical evidence therefore (aswell as paleobotanical data fiom elsewhere in Gondwana) points to the presence of Cretaceous Gondwanan relicts in Madagascar? The break-up of Gondwana by the end of the Cretaceous would have resulted in vicariance in basal angiosperm lineages, and thus, relationships today would be manifest at higher taxonomic levels(VAN STEEMS, 1984). Modern distributions would beessentially austral and/orcool mesic (pre)montane: South America,extendmg northward along the Andes into Central America; Mca, especially cooler upland East Afi-ica (with survival into the Miocene and then extinction in South Africa (COETZEE& MULLER, 1984)); Madagascar; and Australo-malesia east of Wallace's line,i.e., Australia, New Caledonia,New Guinea, and New Zealand. The possiblity of << Noah's Ark )) rafting on India, and therefore the presence of relicts on the granitic islands of the Seychelles and IndidSri Lanka, as well as then subsequent dispersal into W. Malesia must also be considered (a scenariorejected by RAw & AXELROD,1974). Mid to late Cretaceous/earlyTertiary floras from Antarctica, Australia and New Zealand (BIRKENMAJER & ZASTAWNIAK, 1989; BURGER, 1993; DETTMAN, 1989; DETTMAN& JARZEN, 1990; MILDENHALL, 1980)confirm the notion of a subhumidmesothermal climate across Antarctica,which supported a broad-leaved forest withnumerous angiospermpollen types. Spacepermits the discussionof only a fewexamples to illustrate the distribution patterns above: Fig. 1. Schematic representation of Gondwanan land masses during initial Cretaceous radiation of the angiosperms,between the time of arriva1 of'Madagascar atits present position(121 MA) andthe time of separation of GreaterIndia from Madagascar (88 MYA). Throughout the time period, rifting is occurring between both AfiicdSouth America and Greater IndidAntarctica, although each pair remains within close proximity at the end of the period. India is also shown (broken border) after collision with Asia (45 MA), at which point a direct link with W. Malesiais established. Abreviations: MADAG = Madagascar; N.C. = New Caledonia; N.G. = New Guinea (submerged until late Eocene); N.Z. = New Zealand. Proteaceae: The studies of JOHNSONand BRIGG~(198 1) have revealed two independent relict lines of Proteaceae in .Madagascar. Isolated dioecious Dilobeia (2 76 G.E. SCHATZ spp.), the solemember of subtribeDilobeiinae appears to be sister to the subtribe Cenarrheninaeconsisting oftwo monotypicgenera on Tasmania (Agastachys and Cenarrhenes), two genera endemic to New Caledonia (Beaupreopsis (1 sp.) and Beauprea (12 spp.; with fossil distribution in Australia and New Zealand)), and a more xeric taxon Symphionema (2 spp.) in N.S. Wales. From a base number of x = 14 for subfamily Proteoideae, chromosomalevolution indicates aneuploid reductions from Cenarrhenes (x = 14) to Agastachys (x = 13), Beauprea and Beaupreopsis (x = 1l), and S'mphionema (x = lO), as well as polyploidy inDilobeia (x = 24). Subtribe Hickbeachiinae of subfamily Grevilleoideaeis comprised of Malagasia (1 sp.) inMadagascar, Athertonia (1 sp.) and Hicksbeachia (2 spp.) in N.E. Australia, yirotia with 6 spp. probably endemic to New Caledonia (plus 1 sp. doubtllly placed in the genus fromQueensland), and Heliciopsis with 7 spp. restricted to W. Malesia. Relationships among the genera are not fùlly resolved, but if Malagasia is most closely related to Heliciopsis as suggested by JOHNSONand BRIGGS(1981), then one could propose rafting of the ancestors of Heliciopsis on India, dispersal into W. Malesia, and subsequent extinction in India. Basal Apiales: Recent molecular systematic studies by PLUNKETT(1994, pers. comm.) and PLUNKETT et al. (1995) suggest a placement for the Malagasy endemic Melanophylla (8 spp.) at the base of the Apiales among the genera Aralidium (1 sp., W. Malesia), Toricellia (3 spp. E. Himalayas to W. China),and Griselinia (6 spp., New Zealand and Chile), al1 of which have been recognized at the family level. Whether they constitute a clade is still unresolved insofar as several shared indels (insertioddeletion events)within the matK chloroplast gene are plesiomorphicand also shared with outgroup Cornales. Nevertheless, Melanophylla may be closest to Aralidium, reflecting possible rafting on India for their common ancestor as well as for Toricellia, and, in addition, an austral dispersalfor Griselinia [(first appearancein New Zealand in Miocene (MILDENHALL, 1980)], a scenariopotentially analogous to the Hicksbeachiinae (Proteaceae). Monimiaceae: The studies of LORENCE (1985) and PHlLIpsON (1987) have shed light on the relationships of Malagasy region Monimiaceae.Hortonia (2 spp., Sri Lanka), Peumus (1 sp., Chile), Palmeria (14 spp., NewGuinea (mostly), E. Sulawesiand Australia), and Monimia (3 spp., Mascarenes) appear to be the most primitive members of the family (although not necessarily forming a clade), suggesting rafting on Greater India for Hortonia, and extinction of the ancestors of Monzmia in Afik"dagascar given the ages of Mauritius (7.8 My) and Réunion (2 - 3 MY). Decarydendron (3 spp.) and Ephippiandra (8 spp.) (both endemic to Madagascar) and Tambourissa (Ca. 50 spp., Madagascar and Mascarenes) are members of tribe Hedycaryeae, which also includes in E. Malesia: Hedycarya (11 spp., New Caledonia (mostly), New Zealand, Australia to Fiji), Kibaropsis (1 sp., New Caledonia), and Levieria (9 spp., Oueensland, New Guinea to Sulawesi), and in Mica: Xymalos (1-3 spp., from Sudan to S.Africa and E. Congo, disjunct on Mt. Cameroon and Bioko, occurring from 900- 2700 m). Hamamelidaceae: The distribution of subtribe Dicoryphinae mirrors that of the Hedycaryeae (Monimiaceae). Unique tetrasporangiate anthers that dehisce by a single valve per theca constitute a synapomorphy for Trichocladus (3-5 spp., mid-altitude E./S. Mica), Malagasyendemic Dicoryphe (13spp.), and monotypic Noahdendron, Neostp.earia, and Ostrearia in N.E. Australia @NDRESS, 1989a,b). Similarunivalvate anther dehiscence, albeit with a reduction to only one locule per theca is known in PHYTOGEOGRAPHICCONNECTIONS 77 Archamamelis fiom the Upper Cretaceous of Sweden (ENDRESS& FRIIS, 199l), as well as modern-day Hamamelis, indicating an early LaurasidGondwana Split in the family. DidyrnelaceaelSchizocolpus: Recent molecular data suggest that the affinities of the Malagasy endemic Didymeles (Didymelaceae) are with the Buxaceae (M. CHASE, pers. comm.). Distinctive, unique pollen (3-zonocolporate7with 2-orate, operculate colpi anda reticulate, columellate tectum withspinulose muri) has permitted unequivocal .' identification of the fossilpollen Schizocolpusmurlinensis asDidymelaceae. Schizocolpus is known from the Paleocene to Oligocenealong the now-submerged Ninety East Ridge (KEMP& HARRIS, 1977), the Paleocene to lower Eocene of New Zealand (MILDENHALL, 1980)and the middle Eocene GippslandBasin of Australia (STOVER& PARTRIDGE,1973), indicating a formerly much larger late Cretaceous/early Tertiary distribution. Additional probable or possible late Cretaceous/early Tertiary Gondwanan Relicts in Madagascar (or the Mascarenes): Podocarpus (Podocarpaceae); Ascarina (Chloranthaceae); Takhtajania (Winteraceae); some Annonaceae (howevernot an austral dispersal, but rather an early southern Laurasian/N. Tethys dispersal andlor rafting on India);Myristicaceae; Cinnamosma (Canellaceae); Beilschmiedia,Cryptocarya (Lauraceae); IZex (Aquifoliaceae); SyzygiundEugenia (Myrtaceae); Pittosporum (Pittosporaceae); Weinmannia (Cunoniaceae); ElaeocarpzdSloanea (Elaeocarpaceae); Brexia (basal Celastrales); Lilaeopsis (Apiaceae); Protium (?) (Burseraceae) and other basalSapindales, i.e., Rutaceae/Sapindaceae (Cossinia (?) (Sapindaceae) - 1sp. Mauritius,spp. 3 New Caledonia); KaZQhora (?) (Kaliphoraceae);some Menispermaceae (?) (Burasaia,Spirospermum, Strychnopsis); sorneClusiaceae (?) (Garcinia, Symphonia,with Montrouziera in New Caledonia); Arthropodium (Liliaceae); Cohnia (?) (Agavaceae) - 1 sp. Mauritius, 2 spp. New Caledonia; Arecaceae: Borasseae (Borassodendron rafting on India?); Beccariophoenix, Voanioala (Butiinae); Ravenea (Ceroxyloideae); Orania (Oraninae);Iguanurinae (?) (27 genera mainly E. Malesia; absentMadagascar,in Dictyosperma (Mauritius), Bentinckia (S. India)); Oncospermatinae (Oncosperma - Sri Lanka to W. Malesia;absent in Madagascar, 5 genera endemic to Seychelles,2 to Mascarenes); Fandanus (?) (Pandanaceae); some bamboos (?). EOCENE-OLIGOCENE<< LE- STEPPING-STONES >> Calling attention to a distributional pattern that included (Africa)/Madagascar, the Seychelles, (India)/Sri Lanka and W. Malesia, VAN S~EN'IS(1962), in his now infamous paperon the land-bridgetheory, asserted that << there musthave been an isthmian connection between Madagascarand Ceylon over the Seychelles-Comoresbank >) operating during the middle to upper Cretaceous, which he named "Lemuria". Modern understanding of the geological history of the Indian Ocean refbtes both the timing and existence of such directa land route (MCKENZIE & SCLATER, 1973),but the distributional pattern nevertheless still stands, and for some taxa also includes China, Japan,and E. Malesia.Such a pattern often involvesclosely-related species in, and around the periphery of, the Indian Ocean and Africa. As India assumed its current position fiom the early Eocene onward, global sea levels were dropping, with a marked regression at the RupellianKhattian boundary during the .Oligocene, Le., Ca. 30 MYA (HAQ et al., 1988). At that time, significant portions of the ChagoslLaccadive Plateau and the contiguous (at that time) Mascarene Plateau (including the Seychelles Bank, 7s G.E. SCHATZ which encompasses over50,000’km2 now at an average depthof 75 m) could have been emergent, and served as stepping-stones for dispersal of essentially Laurasian (mesic) elements between LaUrasin. Malesia and Africa/Madagascar via IndidSri Lanka as depicted in Figure 2. As such a dispersal track coincides with the << Lemuria )) land- bridge of van Steenis, 1 propose that it be named the <( Lemurian Stepping-stones D. Xericaffinities between Madagascar andIndia (e.g., Commiphora (Burseraceae), Delonix (Fabaceae), Moringa (Moringaceae)) are best explained by overland migration through N.E. AfricdArabian Peninsula, and do not provide additional support for the (( Lemurialand-bridge >) as WILD (1965)suggested. Although the stepping-stone pathway may only have begunoperating during the Eocene, there is no reason to believe that it has not remained active to the present day (Madagascar to Seychelles - 1,000 km; Seychelles to India - 2,600 km), or that migration has not also occurred in an eastward direction [ (e.g., Brexia (Brexiaceae), Grisollea (Icacinaceae)]. Alangium and nacolusa: Paleochorological analysis of Alangiurn (Alangiaceae) and Anacolosa (Olacaceae) (KRUTZSCH, 1989), both originally Laurasian elements that are nowrestricted to Mrica/Madagascar/Asia s.I., supports the stepping-stone hypothesis. Pollenof Anacolosidites (= Anacolosa) and Alangium first appear in India in the upper Eocene, although the former is now absent there. Wlth regard to Anacolosa in Madagascar, the two species differ markedly:A. pervilleana is a deciduous small-leaved, small-fruited species in the dry West, whereas A. casearioides is an evergreen, large- leaved, large-fruited species in humid sublittoralforests along the northeast Coast. Insofar as Anacolosidites is known from the Miocene of east Africa, it is possible the two Malagasy species arrived by different routes. The pollen study of Alangum by REITSMA (1970) suggests a close relationship of dioecious Malagasy A. grisolleoides with A. barbatum in India, and reveals that African A. chinense possesses a distinct and possibly more advanced pollen than the rest of A. chinense, suggesting divergence after an early (Eocene-Oligocene) arriva1 in Africa. The presence of A. salvgolium in the Comores on Mayotte (3.65 - 5.4 MY)and Moheli (1.5 MY)must be a more recent dispersal. Canarium denticulatum Group: The revision of Canarizrm (Burseraceae) by LEENHOUTS(1959) identifies a group of related species ((( denticulatum Group D)that include:endemic species on the Andamans (C. manii), Sri Lanka (C. zeyZanicum), Mauritius (C. paniculatum), as well as C. mahgascariense in Madagascar and E. Africa (hrther taxonomic work is needed in Madagascar to delimit additional species within highlyvariable (( madagascariense D), and C. schweinfurthii in C. and W. Mica. A combination of both primitive and advanced features in C. schweinfurthii suggest long isolation from the rest of the group. On 7.8 MY Mauritius, C. paniculatum probably represents a recent dispersal fi-om Madagascar. Dillenia AmplexicaulGroup: The revisionof Dillenia (Dilleniaceae) by HOOGLAND(1952) recognizes a basic division within the genus on the basis of the presence/absence of anamplexicaul petiolar wing. Both R. triquetra, present in Madagascar and Sri Lanka, and D.ferrugiken, endemic to the Seychelles, possess the winged petiole. Shared distribution of D. triquePa between Madagascar and Sri Lanka implies a more recent dispersal. TheWestern Species of Nepenthes: Paleochorologicalanalysis reveals that Nepenthes was originallya Laurasian element surviving in Europe into the lower Miocene, at the same time as it arrived in Borneo (KRUTZscH, 1989). The small seeds of Nepenthes are undoubtedlyeasily dispersed, and one could envision long distance dispersal fi-omMalesia to Madagascar.Nevertheless, SCHMID-HOLLINGER(1979) has PHYTOGEOGRAPHIC CONNECTIONSPHYTOGEOGRAPHIC 79 suggested that the Western Indian Ocean species of Nepenthes (A? distillatoria - Sri Lanka; A? pewillei - Seychelles; N nzadagascariensis and N masoalensis - Madagascar) form a closely-related group, implying a single initial dispersal eventfiom Malesia. LAURASIA AFRICA - G 1 Fig. 2. Schematic representation of migration pathways active during Eocene/Oligocene after India has attained close proximity to Asia, through to post-collision (and continuing to present),a period during whichsea-levels were dropping. Mesic Laurasian and W. Malesian elements enter Madagascar/Africavia India and {( Lemurian Stepping-Stones )) in the W. Indian Ocean (emergent portions of the Chagos/Laccadive Plateau and Mascarene Plateau including the Seychelles). Xeric elements are exchanged between N.E. AfncdArabian Peninsula and Laurasiahdia and maydisperse to dry western Madagascar. Thicknessof arrow represents probability of dispersal. Additionalpossible examples of LemurianStepping-Stones )) dispersal: Anisophyllea (Anisophylleaceae); Fernandoa (Bignoniaceae);some Clusiaceae (?) - Calophylum; sorne Euphorbiaceae (?) - Claoxylon, Macaranga; Homalium, Scolopia (Flacourtiaceae); Hugonia Wugoniaceae); Foetidiu (Lecythidaceae); Leea (Leeaceae); Medinila, Memecylon (Melastomataceae); Turraea (Meliaceae); Streblus (Moraceae); Ardisia (Myrsinaceae); Paropsia (Passifloraceae); Gaertnera, Ixora, Psychotria, Pyrostria (Rubiaceae). LONGDISTANCE DISPERSAL (LDD) The differencebetween stepping-stone dispersal and long distance dispersalGDD) is only a matter of distance, which isjust one of the factors'contributing to probability of dispersal; the former might also be described as short to medium distance dispersal. Any 80 G.E. SCHATZ dispersal across a fi-agmented space is rendered more probable when a taxon possesses anadded degree of vagility (small, resistant, aerodynamic, or buoyant disseminules), and/or time. That dispersal over especially long distances has occurred is undeniably proven by the many Asian taxa that have reached the Hawaiian Islands in recent history, adistance of 8,000 km. As RAnN (1979)noted, the distancefrom Australia to Madagascar is only about 2/3 (5,400 km) of that fiom Asia to Hawaii, and both source and target have existed for an extremely long time. Prevailing easterly winds and ocean currents in the Indian Oceanfbrther increase the probability for LDD from Malesiato the western Indian Ocean (RENVOISE, 1979).Accepting the fact that LDD ispotentially continuously occurring, impliesthat it has undoubtedly occurred recently, as is witnessed by dispersal to the Young volcanic Mascarenes [(e.g., aphyllodic Acacia ofcertain Australian afE*mity (BELL & EVANS,1978, Who neverthelessconclude a formerland connection between Australia and Mauritius)]. Despite extrinsic directional forces, LDD is often stochastic in nature, and more oRen that not results in a highly imbalanced distribution. Two widespread species of Barringtonia (Lecythidaceae) are among a long listof mangrove and littoral species (the Indo-Pacific Strand flora) to have reached Madagascar by ocean dispersal. Nevertheless, the distributions of B. asiatica and B. racernosa in the western Indian Ocean contain a random component: B. asiatica is present on Mauritius, but not on Réunion, and has failed to reach the E. African Coast; B. racernosa has reached the E. African Coast, but is not present in the Mascarenes; both are present in the Seychelles (PAYENS,1967). To some extent the distributions reflect differing ecological preferences (B. asiatica on sand just above the tide line; B. racernosa in estuarine habitat), but there is also a stochastic element. StrongyZodon Pabaceae) includesa widespread species, S. Zzcidus, thathas reached Hawaii and Tahiti (HIJANG, 1991). In the western Indian Ocean, S. lucidus is present only on Réunion, although Madagascar also harbors a distinct section Craveniae with two species,indicative of an earlierdispersal event. Additional examples of imbalanceddistributions include: GZuta (Anacardiaceae) - 1 sp. Madagascar, 28 spp. Malesia; Hibbertia (Dilleniaceae) - 1(variable) sp. Madagascar, > 100 spp. Malesia, centered in Australia; Keraudrenia (Sterculiaceae) - 1 sp. Madagascar, 8 spp. Australia. CONCLUSIONS As disjunct outliers of the current Indo-australo-malesian flora, those Malagasy humid forest species whose affinitieslie far tothe eastpossess especially high information content, and therefore increased conservation value (VANE-WRIGHTet al., 1991).Barring the discovery of fossilevidence of formerdistribution (COETZEE& MULLER, 1984),extant taxa provide Ouronly rneans of constructinghistorical biogeographic hypotheses. From a phylogenetic standpoint, relict taxa constitute long basalbranches critical for understandingboth ingroup, and broader, higher level outgroup relationships. Extinctions of outlying taxa have greater biogeographic (and possiblyevolutionary) consequences than extinctions of core-area taxa, although,of course,neither is desirable. Just as the Malagasylanguage and customs continue to reveal the S.E Asian origin of the Malagasy people, let us hope that the Malagasy flora will also retainits far eastern character. PHYTOGEOGRAPHC CONNECTIONS 81 ACKNOWLEDGEMENTS 1 wish to dedicate this paper to Professor Hugh H. Iltis, Who fostered my interest in phytogeography. 1 thank A. Le Thomas for translating the abstract, P.P. Lowry II for many stimulatingdiscussions, and the peopleof Madagascar for welcomingvazaha scientistswith such openarms. Work in Madagascarwas carried out through collaborative agreements withthe Parc Botanique et Zoologique de Tsimbazaza and the Département de Recherches Forestières etPiscicoles, FOFIFA. I amgratefùl to the Government of Madagascar (Direction desEaux et Forêts) and the Association National pour la Protection des Aires Protégées for courtesies extended, and to the John D. and Catherine T. MacArthur Foundation and the LizClaiborne and Art Ortenberg Foundation for financial support of activities in Madagascar.This research was conducted while holdingNSF grant DEB-9024749. REFERENCES AUBREVILLE, A., 1976. Madagascar au sein dcla Pangée. Adansonia, sér. 2, 15: 295-305. AXELROD, D.I. & P.H. RAVEN, 1982. Paleobiogeographyand origin of the New Guinea flora. In: J.L. Gressit (ed.), Biogeography and Ecologyof New Guinea. pp. 919-941. W. Junk, The Hague. BELL, E.A. & CS. EVANS, 1978. Biochemical evidence of a former link between Australia and the Mascarene Islands. Nature, 273: 295-296. BIRKENMAJER K. & E. ZASTAWNIAK, 1989. Late Cretaceous-Early Tertiary floras of King George Island, West Antarctica: their stratigraphic distribution and palaeoclimatic sigmfïcance. In: J.A.Crame (ed.), Origins and Evolution of the AntarcticBiota. pp. 227-240.Geol. Soc. Special Publ., 47, London. BURGER D., 1993.Early and middle Cretaceous angiosperm pollen grains from Australia. Rev. Palaeobot. Palynol., 73: 183-234. COETZEE, J.A. & J. MULLER 1984. The phytogeographic sigmficance of some extinct Gondwana pollen types from the Tertiary of the smthwestern Cape (South Africa). Ann. Missouri Bot. Gad., 71: 1088-1099. DEJARDIN,J., J.-L. GUILLAUMET & G. hfANGENOT, 1973.Contribution à laconnaissance de l'élément non endémique deLa flore malgache (végétaux vasculaires). Candolleaj28: 325-391. DETTMAN, M.E., 1989. Antarctica: Cretaceous cradleof austral temperate rainforests? In: Crame, J.A. (ed.), Origins and Evolution of the Antarctic Biota. pp.89-105, Geol. Soc. Special. Publ., 47, London. DETTMAN, M.E. & D.M. JARZEN, 1990. The AntarctidAustralian rift Valley: Late Cretaceous cradle of northeastern Australasian relicts?Rev. Palaeobot. Palynol., 65: 13 1-144. ENDRESS, P.K., 1989a. Aspects of evolutionary differenliation of the Hamamelidaceae and the lower Hamamelididae. Pl. Syst. Evol.,. 162: 193-211. 82 SCHATZ G.E. ENDKESS,P.K., 1989c. Phylogeneticrelationships in Hamamelidoideae. In: Crane, P.R. & S. Blackmore (eds.), Evolution, Systematics and FossilHistory of the Hamamelidae. Vol. 1. Introductionand (< Lower D Hamamelidae.pp. 227-248. SystematicsAssociation Special Volume No. 40A. Clarendon Press, Oxford. ENDRESS, P.K. & E.M.FRIIS, 1991. Archumamelis: hamamelidaleanflowers from the Upper Creataceous of Sweden. Pl. Syst. Evol., 175: 101-114. HAQ,B.U., J. HARDEMOL & P.R.VAIL, 1988. Mesozoicand Cenozoic chronostratigraphy and eustatic cycles. In: C.K. Wilgus, B.S. Hastings, H. Posamentier, J.V. Wagoner,C.A. Ross and C.G.StC. Kendall (eds.), Sea-level changes:an integrated approach. pp.71-108. Special Publ. 42, Soc. Econ. Paleont. Mineral., Tulsa, Oklahoma. RERNGREEN, G.F.W., A. RANDRIANASOLO& J.W. VERBEEK, 1982. Micropaleontology of Albian to Dania stratain Madagascar. Micropaleontology,28: 97-109. HOOGLAND, R.D., 1952. A revisionof the genus Dilleniu. Blumea, 7: 1-145. HUANG, S.-F., 199 1. StrongyIodon (Leguminosae-Erythtlnae), a revision of the genus. Wageningen Agric. U. papers: 90-8. HUMBERT, H., 1959. Originespresumées et affinitésde la florede Madagascar. Mém. Inst. Sci. Madagascar, sér.B, 9: 149-187. JOHNSON, L.A.S. & B.G. BRIGGS, 1981. Three old southern families: Myrtaceae, Proteaceae, and Restionaceae. In: A. Keast (ed.), Ecological Biogeography of Australia. pp.427-469, W. Junk, The Hague. KEMP, E.M. & W.K. HARRIS, 1975. The vegetation of Tertiary islands on the Ninety East Ridge. Nature, 258: 303-307. KOECHLIN,J., J.-L. GUILLAUMET & P.MORAT, 1972. Floreet Végétation de Madagascar. J. Cramer, Berlin.687p. KRUTZSCH,W., 1989. Paleogeography and historicalphytogeography (paleochorology) in the Neophyticum. Pl. Syst. Evol., 162: 5-61. LEENHOUTS, P.W., 1959. Revision of the Burseraceaeof the Malaysia areain a wider sense. Blumea, 9: 275-475. LEROY,J.-F., 1978. Composition,origin, and aflïnities of theMadagascan vascular flora. Ann. Missouri Bot. Gard., 65: 535-589. LORENCE,D.H., 1985. A monograph of the Monimiaceae(Laurales) in the MalagasyRegion (Southwest Indian Ocean).Ann. Missouri Bot. Gard., 72: 1-165. MCKENZIE, D.P. & J.C. SCLAEK 1973. The evolution of the Indian Ocean. Sci. American, 228(5): 63-72. MILDENHALL, D.C., 1980. New ZealandLate Cretaceous and Cenozoic plant biogeography:a contribution. Palaeogeogr. Palaeoclimatol. Palaeoecol.,3 1: 197-233. PAYENS, J.P.D.W., 1967. A monograph of the genus Burringfonia (Lecythidaceae). Blumea, 15: 157- 263. PERRIER DE LA BA=, H., 1936. Biogéographie des Plantes de Madagascar. Société d'Editions, Géographiques, Maritimes et Coloniales, Paris. PHILIF'SON, W.R., 1987. A classificationof the Monimiaceae. NordicJ. Bot., 7: 25-29, PHYTO GEO GRA PHIC CONNECTIONSPHYTOGEOGRAPHIC 83 PLUNKETT, G.M., 1994. A molecular-cladistic approach tothe (( family-pair dilemman in Apiales and Cyperales. Ph.D. Dissertation. Washington State University, Pullman, Washington,173 pp. PLUNKETT, G.M., D.E. SOLTIS & P.S. SOLTIS, 1995. Higher level relationships involving Apiales (Apiaceae and Araliaceae) based on phylogenetic analysisof rbcL sequences. AmerJ. Bot. (in press). RABINOWITZ, PD., M.F. COFFIN & D. FALVEY, 1983. The separation of Madagascar and Bca. Science, 220: 67-69. RAVEN,P.H., 1979. Platetectonics and Southem Hemisphere biogeography. In: K. Larsen(ed.), Tropical Botany. pp. 1-24. Academic Press, London. RAVEN, P.H. & D.I. AXELROD, 1974. Angiosperm biogeography and past continental movements. Ann. Missouri Bot. Gard., 61: 539-673. REITSMA, T., 1970. Pollen morphologyof the Alangiaceae.Rev. Palaeobot. Palynol., 10: 249-332. RENVOISE, S.A., 1979. The origins of Indian Ocean island floras. In: D. Bramwell (ed.), Plants and Islands. pp. 107-129. Academic Press, London. SCHMID-HOLLINGER, R, 1979. Nepenthes-Studien V. Die Kannenformen der westlichenNepenthes- Arten. Bot. Jahrb. Syst., 100: 379-405. STEENIS? C.G.G.J. van,1962. The land-bridge theoryin botany. Blumea, 11: 235-372. STEENIS, C.G.G.J. van, 1984. Plant geographyof east Malesia. Bot. J. Linn. Soc., 79: 97-178. STOREY M., J. J. MAHONEY, A.D. SAUNDERS, R. A. DUNCAN, S.P KELLEY, & M.F. COWIN, 1995. Timing of hotspot-related volcanism and the breakup of Madagascarand India. Science, 267: 852-855. STOVER, L.E. & A.D. PARTRIDGE, 1973. Tertiary and Late Cretaceous spores and pollen from the Gippsland Basin, Southeastem Australia. Proc. R. Soc. Vkt.,85: 237-286. VANE-WRIGHT, R.I., C.J. HUMPHRlES & P.H. WILLIAMS, 1991. What to protect? - Systematics and the agony of choice. Biol. Conservation,55: 235-254. WILD, H., 1965. Additional evidence for the Afiica-Madagascar-India-Ceylon land-bridge theory with special reference tothe genera Anisopappus and Cornrniphora.Webbia, 19: 497-505.